relying on open models, each of them communicating through published APIs.
-.. figure:: ./images/TransportPCE-Diagramm-Magnesium.jpg
+.. figure:: ./images/TransportPCE-Diagram-Phosphorus.jpg
:alt: TransportPCE architecture
TransportPCE architecture
and transponders.
The interest of using a controller to provision automatically services strongly
-relies on its ability to handle end to end optical services that spans through
+relies on its ability to handle end to end optical services that span through
the different network domains, potentially equipped with equipment coming from
different suppliers. Thus, interoperability in the optical layer is a key
element to get the benefit of automated control.
interoperability and Yang data models.
End to end OTN services such as OCH-OTU4, structured ODU4 or 10GE-ODU2e
-services are supported since Magnesium SR2. OTN support will continue to be
+services are supported since Magnesium SR2. OTN support continued to be
improved in the following releases of Magnesium and Aluminium.
-An experimental support of Flexgrid is introduced in Aluminium. Depending on
-OpenROADM device models, optical interfaces can be created according to the
-initial fixed grid (for R1.2.1, 96 channels regularly spaced of 50 GHz), or to
-a flexgrid (for R2.2.1 use of specific number of subsequent frequency slots of
-6.25 GHz depending on one side of ROADMs and transponders capabilities and on
-the other side of the rate of the channel. The full support of Flexgrid,
-including path computation and the creation of B100G (Beyond 100 Gbps) higher
-rate interfaces will be added in the following releases of Aluminium.
+Flexgrid was introduced in Aluminium. Depending on OpenROADM device models,
+optical interfaces can be created according to the initial fixed grid (for
+R1.2.1, 96 channels regularly spaced of 50 GHz), or to a flexgrid (for R2.2.1
+use of specific number of subsequent frequency slots of 6.25 GHz depending on
+one side of ROADMs and transponders capabilities and on the other side of the
+rate of the channel.
+
+Leveraging Flexgrid feature, high rate services are supported since Silicon.
+First implementation allows rendering 400 GE services. This release also brings
+asynchronous service creation and deletion, thanks to northbound notifications
+modules based on a Kafka implementation, allowing interactions with the DMaaP
+Bus of ONAP.
+
+Phosphorus consolidates end to end support for high rate services (ODUC4, OTUC4),
+allowing service creation and deletion from the NBI. The support of path
+computation for high rate services (OTUC4) will be added through the different P
+releases, relying on GNPy for impairment aware path computation. An experimental
+support of T-API is provided allowing service-create/delete from a T-API version
+2.1.1 compliant NBI. A T-API network topology, with different levels of abstraction
+and service context are maintained in the MDSAL. Service state is managed,
+monitoring device port state changes. Associated notifications are handled through
+Kafka and DMaaP clients.
Module description
ServiceHandler
^^^^^^^^^^^^^^
-Service Handler handles request coming from a higher level controller or an orchestrator
-through the northbound API, as defined in the Open ROADM service model. Current
-implementation addresses the following rpcs: service-create, temp-service-create,
-service–delete, temp-service-delete, service-reroute, and service-restoration. It checks the
-request consistency and trigs path calculation sending rpcs to the PCE. If a valid path is
-returned by the PCE, path configuration is initiated relying on Renderer and OLM. At the
-confirmation of a successful service creation, the Service Handler updates the service-
-list/temp-service-list in the MD-SAL. For service deletion, the Service Handler relies on the
-Renderer and the OLM to delete connections and reset power levels associated with the
-service. The service-list is updated following a successful service deletion. In Neon SR0 is
-added the support for service from ROADM to ROADM, which brings additional flexibility and
-notably allows reserving resources when transponders are not in place at day one.
-Magnesium SR2 fully supports end-to-end OTN services which are part of the OTN infrastructure.
-It concerns the management of OCH-OTU4 (also part of the optical infrastructure) and structured
-HO-ODU4 services. Moreover, once these two kinds of OTN infrastructure service created, it is
-possible to manage some LO-ODU services (for the time being, only 10GE-ODU2e services).
-The full support of OTN services, including 1GE-ODU0 or 100GE, will be introduced along next
-releases (Mg/Al).
+Service Handler handles request coming from a higher level controller or an
+orchestrator through the northbound API, as defined in the Open ROADM service model.
+Current implementation addresses the following rpcs: service-create, temp-service-
+create, service–delete, temp-service-delete, service-reroute, and service-restoration.
+It checks the request consistency and trigs path calculation sending rpcs to the PCE.
+If a valid path is returned by the PCE, path configuration is initiated relying on
+Renderer and OLM. At the confirmation of a successful service creation, the Service
+Handler updates the service-list/temp-service-list in the MD-SAL. For service deletion,
+the Service Handler relies on the Renderer and the OLM to delete connections and reset
+power levels associated with the service. The service-list is updated following a
+successful service deletion. In Neon SR0 is added the support for service from ROADM
+to ROADM, which brings additional flexibility and notably allows reserving resources
+when transponders are not in place at day one. Magnesium SR2 fully supports end-to-end
+OTN services which are part of the OTN infrastructure. It concerns the management of
+OCH-OTU4 (also part of the optical infrastructure) and structured HO-ODU4 services.
+Moreover, once these two kinds of OTN infrastructure service created, it is possible
+to manage some LO-ODU services (1GE-ODU0, 10GE-ODU2e). 100GE services are also
+supported over ODU4 in transponders or switchponders using higher rate network
+interfaces.
+
+In Silicon release, the management of TopologyUpdateNotification coming from the *Topology Management*
+module was implemented. This functionality enables the controller to update the information of existing
+services according to the online status of the network infrastructure. If any service is affected by
+the topology update and the *odl-transportpce-nbi* feature is installed, the Service Handler will send a
+notification to a Kafka server with the service update information.
PCE
^^^
In Magnesium SR0, the interconnection of the PCE with GNPY (Gaussian Noise Python), an
open-source library developed in the scope of the Telecom Infra Project for building route
-planning and optimizing performance in optical mesh networks, is fully supported.
+planning and optimizing performance in optical mesh networks, is fully supported. Impairment
+aware path computation for service of higher rates (Beyond 100G) is planned across Phoshorus
+releases. It implies to make B100G OpenROADM specifications available in GNPy libraries.
If the OSNR calculated by the PCE is too close to the limit defined in OpenROADM
specifications, the PCE forwards through a REST interface to GNPY external tool the topology
The result is sent back to the PCE. If the path is validated, the PCE sends back a response to
the service handler. In case of invalidation of the path by GNPY, the PCE sends a new request to
GNPY, including only the constraints expressed in the path-computation-request initiated by the
-Service Handler. GNPy then tries to calculate a path based on these relaxed constraints. The result
-of the path computation is provided to the PCE which translates the path according to the topology
-handled in transportPCE and forwards the results to the Service Handler.
+Service Handler. GNPy then tries to calculate a path based on these relaxed constraints. The
+result of the path computation is provided to the PCE which translates the path according to the
+topology handled in transportPCE and forwards the results to the Service Handler.
GNPy relies on SNR and takes into account the linear and non-linear impairments
to check feasibility. In the related tests, GNPy module runs externally in a
The population of OTN links (OTU4 and ODU4), and the adjustment of the tributary ports/slots
pool occupancy when OTN services are created is supported since Magnesium SR2.**
+Since Silicon release, the Topology Management module process NETCONF event received through an
+event stream (as defined in RFC 5277) between devices and the NETCONF adapter of the controller.
+Current implementation detects device configuration changes and updates the topology datastore accordingly.
+Then, it sends a TopologyUpdateNotification to the *Service Handler* to indicate that a change has been
+detected in the network that may affect some of the already existing services.
Renderer
^^^^^^^^
device models. The full support of openroadm 2.2.1 device models (both in the topology management
and the rendering function) has been added in Neon SR1. In Magnesium, portmapping is enriched with
the supported-interface-capability, OTN supporting-interfaces, and switching-pools (reflecting
-cross-connection capabilities of OTN switch-ponders).
+cross-connection capabilities of OTN switch-ponders). The support for 7.1 devices models is
+introduced in Silicon (no devices of intermediate releases have been proposed and made available
+to the market by equipment manufacturers).
After the path is provided, the renderer first checks what are the existing interfaces on the
ports of the different nodes that the path crosses. It then creates missing interfaces. After all
Magnesium brings the support of OTN services. SR0 supports the creation of OTU4, ODU4, ODU2/ODU2e
and ODU0 interfaces. The creation of these low-order otn interfaces must be triggered through
otn-service-path RPC. Magnesium SR2 fully supports end-to-end otn service implementation into devices
-(service-implementation-request /service delete rpc, topology alignement after the service has been created).
+(service-implementation-request /service delete rpc, topology alignement after the service
+has been created).
+In Silicon releases, higher rate OTN interfaces (OTUC4) must be triggered through otn-service-
+path RPC. Phosphorus SR0 supports end-to-end otn service implementation into devices
+(service-implementation-request /service delete rpc, topology alignement after the service
+has been created). One shall note that impairment aware path calculation for higher rates will
+be made available across the Phosphorus release train.
OLM
^^^
Inventory
^^^^^^^^^
-TransportPCE Inventory module is responsible to keep track of devices connected in an external MariaDB database.
-Other databases may be used as long as they comply with SQL and are compatible with OpenDaylight (for example MySQL).
-At present, the module supports extracting and persisting inventory of devices OpenROADM MSA version 1.2.1.
-Inventory module changes to support newer device models (2.2.1, etc) and other models (network, service, etc)
-will be progressively included.
+TransportPCE Inventory module is responsible to keep track of devices connected in an external
+MariaDB database. Other databases may be used as long as they comply with SQL and are compatible
+with OpenDaylight (for example MySQL). At present, the module supports extracting and persisting
+inventory of devices OpenROADM MSA version 1.2.1. Inventory module changes to support newer device
+models (2.2.1, etc) and other models (network, service, etc) will be progressively included.
The inventory module can be activated by the associated karaf feature (odl-transporpce-inventory)
-The database properties are supplied in the “opendaylight-release” and “opendaylight-snapshots” profiles.
-Below is the settings.xml with properties included in the distribution.
+The database properties are supplied in the “opendaylight-release” and “opendaylight-snapshots”
+profiles. Below is the settings.xml with properties included in the distribution.
The module can be rebuild from sources with different parameters.
Sample entry in settings.xml to declare an external inventory database:
</profiles>
-Once the project built and when karaf is started, the cfg file is generated in etc folder with the corresponding
-properties supplied in settings.xml. When devices with OpenROADM 1.2.1 device model are mounted, the device listener in
-the inventory module loads several device attributes to various tables as per the supplied database.
-The database structure details can be retrieved from the file tests/inventory/initdb.sql inside project sources.
-Installation scripts and a docker file are also provided.
+Once the project built and when karaf is started, the cfg file is generated in etc folder with the
+corresponding properties supplied in settings.xml. When devices with OpenROADM 1.2.1 device model
+are mounted, the device listener in the inventory module loads several device attributes to various
+tables as per the supplied database. The database structure details can be retrieved from the file
+tests/inventory/initdb.sql inside project sources. Installation scripts and a docker file are also
+provided.
Key APIs and Interfaces
-----------------------
- PCE to Topology Management
- Service Handler to Renderer
- Renderer to OLM
+- Network Model to Service Handler
Pce Service
^^^^^^^^^^^
- This feature provides function to be able to stub some of TransportPCE modules, pce and
renderer (Stubpce and Stubrenderer).
- Stubs are used for development purposes and can be used for some of the functionnal tests.
+ Stubs are used for development purposes and can be used for some of the functional tests.
Interfaces to external software
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Connecting nodes
~~~~~~~~~~~~~~~~
-To connect a node, use the following JSON RPC
+To connect a node, use the following RESTconf request
-**REST API** : *POST /restconf/config/network-topology:network-topology/topology/topology-netconf/node/<node-id>*
+**REST API** : *PUT /rests/data/network-topology:network-topology/topology=topology-netconf/node=<node-id>*
**Sample JSON Data**
Then check that the netconf session has been correctly established between the controller and the
node. the status of **netconf-node-topology:connection-status** must be **connected**
-**REST API** : *GET /restconf/operational/network-topology:network-topology/topology/topology-netconf/node/<node-id>*
+**REST API** : *GET /rests/data/network-topology:network-topology/topology=topology-netconf/node=<node-id>?content=nonconfig*
Node configuration discovery
physical ports related to transmission. All *circuit-packs* inside the node configuration are
analyzed.
-Use the following JSON RPC to check that function internally named *portMapping*.
+Use the following RESTconf URI to check that function internally named *portMapping*.
-**REST API** : *GET /restconf/config/portmapping:network*
+**REST API** : *GET /rests/data/transportpce-portmapping:network*
.. note::
topology can be partial. Check that link of type *ROADMtoROADM* exists between two adjacent rdm
nodes.
-**REST API** : *GET /restconf/config/ietf-network:network/openroadm-topology*
+**REST API** : *GET /rests/data/ietf-network:networks/network=openroadm-topology*
If it is not the case, you need to manually complement the topology with *ROADMtoROADM* link using
the following REST RPC:
-**REST API** : *POST /restconf/operations/networkutils:init-roadm-nodes*
+**REST API** : *POST /rests/operations/transportpce-networkutils:init-roadm-nodes*
**Sample JSON Data**
.. code:: json
{
- "networkutils:input": {
- "networkutils:rdm-a-node": "<node-id-A>",
- "networkutils:deg-a-num": "<degree-A-number>",
- "networkutils:termination-point-a": "<Logical-Connection-Point>",
- "networkutils:rdm-z-node": "<node-id-Z>",
- "networkutils:deg-z-num": "<degree-Z-number>",
- "networkutils:termination-point-z": "<Logical-Connection-Point>"
+ "input": {
+ "rdm-a-node": "<node-id-A>",
+ "deg-a-num": "<degree-A-number>",
+ "termination-point-a": "<Logical-Connection-Point>",
+ "rdm-z-node": "<node-id-Z>",
+ "deg-z-num": "<degree-Z-number>",
+ "termination-point-z": "<Logical-Connection-Point>"
}
}
From xpdr to rdm:
^^^^^^^^^^^^^^^^^
-**REST API** : *POST /restconf/operations/networkutils:init-xpdr-rdm-links*
+**REST API** : *POST /rests/operations/transportpce-networkutils:init-xpdr-rdm-links*
**Sample JSON Data**
.. code:: json
{
- "networkutils:input": {
- "networkutils:links-input": {
- "networkutils:xpdr-node": "<xpdr-node-id>",
- "networkutils:xpdr-num": "1",
- "networkutils:network-num": "<xpdr-network-port-number>",
- "networkutils:rdm-node": "<rdm-node-id>",
- "networkutils:srg-num": "<srg-number>",
- "networkutils:termination-point-num": "<Logical-Connection-Point>"
+ "input": {
+ "links-input": {
+ "xpdr-node": "<xpdr-node-id>",
+ "xpdr-num": "1",
+ "network-num": "<xpdr-network-port-number>",
+ "rdm-node": "<rdm-node-id>",
+ "srg-num": "<srg-number>",
+ "termination-point-num": "<Logical-Connection-Point>"
}
}
}
From rdm to xpdr:
^^^^^^^^^^^^^^^^^
-**REST API** : *POST /restconf/operations/networkutils:init-rdm-xpdr-links*
+**REST API** : *POST /rests/operations/transportpce-networkutils:init-rdm-xpdr-links*
**Sample JSON Data**
.. code:: json
{
- "networkutils:input": {
- "networkutils:links-input": {
- "networkutils:xpdr-node": "<xpdr-node-id>",
- "networkutils:xpdr-num": "1",
- "networkutils:network-num": "<xpdr-network-port-number>",
- "networkutils:rdm-node": "<rdm-node-id>",
- "networkutils:srg-num": "<srg-number>",
- "networkutils:termination-point-num": "<Logical-Connection-Point>"
+ "input": {
+ "links-input": {
+ "xpdr-node": "<xpdr-node-id>",
+ "xpdr-num": "1",
+ "network-num": "<xpdr-network-port-number>",
+ "rdm-node": "<rdm-node-id>",
+ "srg-num": "<srg-number>",
+ "termination-point-num": "<Logical-Connection-Point>"
}
}
}
or SWITCH type connected to two different rdm devices. To check that these xpdr are present in the
OTN topology, use the following command on the REST API :
-**REST API** : *GET /restconf/config/ietf-network:network/otn-topology*
+**REST API** : *GET /rests/data/ietf-network:networks/network=otn-topology*
An optical connectivity service shall have been created in a first setp. Since Magnesium SR2, the OTN
links are automatically populated in the topology after the Och, OTU4 and ODU4 interfaces have
~~~~~~~~~~~~~~~~~~
Use the *service handler* module to create any end-to-end connectivity service on an OpenROADM
-network. Two kind of end-to-end "optical" services are managed by TransportPCE:
-- 100GE service from client port to client port of two transponders (TPDR)
+network. Two different kinds of end-to-end "optical" services are managed by TransportPCE:
+- 100GE/400GE services from client port to client port of two transponders (TPDR)
- Optical Channel (OC) service from client add/drop port (PP port of SRG) to client add/drop port of
two ROADMs.
For these services, TransportPCE automatically invokes *renderer* module to create all required
interfaces and cross-connection on each device supporting the service.
-As an example, the creation of a 100GE service implies among other things, the creation of OCH, OTU4
-and ODU4 interfaces on the Network port of TPDR devices.
+As an example, the creation of a 100GE service implies among other things, the creation of OCH or
+Optical Tributary Signal (OTSi), OTU4 and ODU4 interfaces on the Network port of TPDR devices.
+The creation of a 400GE service implies the creation of OTSi, OTUC4, ODUC4 and ODU4 interfaces on
+the Network port of TPDR devices.
Since Magnesium SR2, the *service handler* module directly manages some end-to-end otn
connectivity services.
*openroadm-topology* and then invokes *renderer* and *OLM* to implement the end-to-end path into
the devices.
+OTSi-OTUC4 service creation
+^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Use the following REST RPC to invoke *service handler* module in order to create over the optical
+infrastructure a bidirectional end-to-end OTUC4 over an optical Optical Tributary Signal
+connectivity service between two optical network ports of OTN Xponder (MUXPDR or SWITCH). Such
+service configure the optical network infrastructure composed of rdm nodes.
+
+**REST API** : *POST /restconf/operations/org-openroadm-service:service-create*
+
+**Sample JSON Data**
+
+.. code:: json
+
+ {
+ "input": {
+ "sdnc-request-header": {
+ "request-id": "request-1",
+ "rpc-action": "service-create",
+ "request-system-id": "appname"
+ },
+ "service-name": "something",
+ "common-id": "commonId",
+ "connection-type": "infrastructure",
+ "service-a-end": {
+ "service-rate": "400",
+ "node-id": "<xpdr-node-id>",
+ "service-format": "OTU",
+ "otu-service-rate": "org-openroadm-otn-common-types:OTUCn",
+ "clli": "<ccli-name>",
+ "tx-direction": {
+ "port": {
+ "port-device-name": "<xpdr-node-id-in-otn-topology>",
+ "port-type": "fixed",
+ "port-name": "<xpdr-network-port-in-otn-topology>",
+ "port-rack": "000000.00",
+ "port-shelf": "Chassis#1"
+ },
+ "lgx": {
+ "lgx-device-name": "Some lgx-device-name",
+ "lgx-port-name": "Some lgx-port-name",
+ "lgx-port-rack": "000000.00",
+ "lgx-port-shelf": "00"
+ }
+ },
+ "rx-direction": {
+ "port": {
+ "port-device-name": "<xpdr-node-id-in-otn-topology>",
+ "port-type": "fixed",
+ "port-name": "<xpdr-network-port-in-otn-topology>",
+ "port-rack": "000000.00",
+ "port-shelf": "Chassis#1"
+ },
+ "lgx": {
+ "lgx-device-name": "Some lgx-device-name",
+ "lgx-port-name": "Some lgx-port-name",
+ "lgx-port-rack": "000000.00",
+ "lgx-port-shelf": "00"
+ }
+ },
+ "optic-type": "gray"
+ },
+ "service-z-end": {
+ "service-rate": "400",
+ "node-id": "<xpdr-node-id>",
+ "service-format": "OTU",
+ "otu-service-rate": "org-openroadm-otn-common-types:OTUCn",
+ "clli": "<ccli-name>",
+ "tx-direction": {
+ "port": {
+ "port-device-name": "<xpdr-node-id-in-otn-topology>",
+ "port-type": "fixed",
+ "port-name": "<xpdr-network-port-in-otn-topology>",
+ "port-rack": "000000.00",
+ "port-shelf": "Chassis#1"
+ },
+ "lgx": {
+ "lgx-device-name": "Some lgx-device-name",
+ "lgx-port-name": "Some lgx-port-name",
+ "lgx-port-rack": "000000.00",
+ "lgx-port-shelf": "00"
+ }
+ },
+ "rx-direction": {
+ "port": {
+ "port-device-name": "<xpdr-node-id-in-otn-topology>",
+ "port-type": "fixed",
+ "port-name": "<xpdr-network-port-in-otn-topology>",
+ "port-rack": "000000.00",
+ "port-shelf": "Chassis#1"
+ },
+ "lgx": {
+ "lgx-device-name": "Some lgx-device-name",
+ "lgx-port-name": "Some lgx-port-name",
+ "lgx-port-rack": "000000.00",
+ "lgx-port-shelf": "00"
+ }
+ },
+ "optic-type": "gray"
+ },
+ "due-date": "yyyy-mm-ddT00:00:01Z",
+ "operator-contact": "some-contact-info"
+ }
+ }
+
+As for the previous RPC, this RPC invokes the *PCE* module to compute a path over the
+*openroadm-topology* and then invokes *renderer* and *OLM* to implement the end-to-end path into
+the devices.
+
+One shall note that in Phosphorus SR0, as the OpenROADM 400G specification are not available (neither
+in the GNPy libraries, nor in the *PCE* module), path validation will be performed using the same
+asumptions as we use for 100G. This means the path may be validated whereas optical performances do
+not reach expected levels. This allows testing OpenROADM device implementing B100G rates, but shall
+not be used in operational conditions. The support for higher rate impairment aware path computation
+will be introduced across Phosphorus release train.
+
+ODUC4 service creation
+^^^^^^^^^^^^^^^^^^^^^^
+
+For ODUC4 service creation, the REST RPC to invoke *service handler* module in order to create an
+ODUC4 over the OTSi-OTUC4 has the same format as the RPC used for the creation of this last. Only
+"service-format" needs to be changed to "ODU", and "otu-service-rate" : "org-openroadm-otn-common-
+types:OTUCn" needs to be replaced by: "odu-service-rate" : "org-openroadm-otn-common-types:ODUCn"
+in both service-a-end and service-z-end containers.
+
OTN HO-ODU4 service creation
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Magnesium SR0). The trib-slot provided corresponds to the first of the used trib-slots.
complex-trib-slots will be used when the equipment does not support contiguous trib-slot
allocation. In this case a list of the different trib-slots to be used shall be provided.
- The support for non contiguous trib-slot allocation is planned for later Magnesium release.
+ The support for non contiguous trib-slot allocation is planned for later release.
Deleting a service
~~~~~~~~~~~~~~~~~~
here, the <otn-node-id> corresponds to the node-id as appearing in "otn-topology" layer
+odl-transportpce-tapi
+---------------------
+
+This feature allows TransportPCE application to expose at its northbound interface other APIs than
+those defined by the OpenROADM MSA. With this feature, TransportPCE provides part of the Transport-API
+specified by the Open Networking Foundation. More specifically, the Topology Service and Connectivity
+Service components are implemented, allowing to expose to higher level applications an abstraction of
+its OpenROADM topologies in the form of topologies respecting the T-API modelling, as well as
+creating/deleting connectivity services between the Service Interface Points (SIPs) exposed by the
+T-API topology. The current version of TransportPCE implements the *tapi-topology.yang* and
+*tapi-connectivity.yang* models in the revision 2018-12-10 (T-API v2.1.2).
+
+Additionally, support for the Notification Service component will be added in future releases, which
+will allow higher level applications to create/delete a Notification Subscription Service to receive
+several T-API notifications as defined in the *tapi-notification.yang* model.
+
+T-API Topology Service
+~~~~~~~~~~~~~~~~~~~~~~
+
+- RPC calls implemented:
+
+ - get-topology-details
+
+ - get-node-details
+
+ - get-node-edge-point-details
+
+ - get-link-details
+
+ - get-topology-list
+
+
+As in IETF or OpenROADM topologies, T-API topologies are composed of lists of nodes and links that
+abstract a set of network resources. T-API specifies the *T0 - Multi-layer topology* which is, as
+indicated by its name, a single topology that collapses network logical abstraction for all network
+layers. Thus, an OpenROADM device as, for example, an OTN xponder that manages the following network
+layers ETH, ODU, OTU, Optical wavelength, will be represented in T-API T0 topology by two nodes:
+one *DSR/ODU* node and one *Photonic Media* node. Each of them are linked together through one or
+several *transitional links* depending on the number of network/line ports on the device.
+
+Aluminium SR2 comes with a complete refactoring of this module, handling the same way multi-layer
+abstraction of any Xponder terminal device, whether it is a 100G transponder, an OTN muxponder or
+again an OTN switch. For all these devices, the implementation manages the fact that only relevant
+ports must appear in the resulting TAPI topology abstraction. In other words, only client/network ports
+that are undirectly/directly connected to the ROADM infrastructure are considered for the abstraction.
+Moreover, the whole ROADM infrastructure of the network is also abstracted towards a single photonic
+node. Therefore, a pair of unidirectional xponder-output/xponder-input links present in *openroadm-topology*
+is represented by a bidirectional *OMS* link in TAPI topology.
+In the same way, a pair of unidirectional OTN links (OTU4, ODU4) present in *otn-topology* is also
+represented by a bidirectional OTN link in TAPI topology, while retaining their available bandwidth
+characteristics.
+
+Phosphorus SR0 extends the T-API topology service implementation by bringing a fully described topology.
+*T0 - Full Multi-layer topology* is derived from the existing *T0 - Multi-layer topology*. But the ROADM
+infrastructure is not abstracted and the higher level application can get more details on the composition
+of the ROADM infrastructure controlled by TransportPCE. Each ROADM node found in the *openroadm-network*
+is converted into a *Photonic Media* node. The details of these T-API nodes are obtained from the
+*openroadm-topology*. Therefore, the external traffic ports of *Degree* and *SRG* nodes are represented
+with a set of Network Edge Points (NEPs) and SIPs belonging to the *Photonic Media* node and a pair of
+roadm-to-roadm links present in *openroadm-topology* is represented by a bidirectional *OMS* link in TAPI
+topology.
+Additionally, T-API topology related information is stored in TransportPCE datastore in the same way as
+OpenROADM topology layers. When a node is connected to the controller through the corresponding *REST API*,
+the T-API topology context gets updated dynamically and stored.
+
+.. note::
+
+ A naming nomenclature is defined to be able to map T-API and OpenROADM data.
+ i.e., T-API_roadm_Name = OpenROADM_roadmID+T-API_layer
+ i.e., T-API_roadm_nep_Name = OpenROADM_roadmID+T-API_layer+OpenROADM_terminationPointID
+
+Three kinds of topologies are currently implemented. The first one is the *"T0 - Multi-layer topology"*
+defined in the reference implementation of T-API. This topology gives an abstraction from data coming
+from openroadm-topology and otn-topology. Such topology may be rather complex since most of devices are
+represented through several nodes and links.
+Another topology, named *"Transponder 100GE"*, is also implemented. That latter provides a higher level
+of abstraction, much simpler, for the specific case of 100GE transponder, in the form of a single
+DSR node.
+Lastly, the *T0 - Full Multi-layer topology* topology was added. This topology collapses the data coming
+from openroadm-network, openroadm-topology and otn-topology. It gives a complete view of the optical
+network as defined in the reference implementation of T-API
+
+The figure below shows an example of TAPI abstractions as performed by TransportPCE starting from Aluminium SR2.
+
+.. figure:: ./images/TransportPCE-tapi-abstraction.jpg
+ :alt: Example of T0-multi-layer TAPI abstraction in TransportPCE
+
+In this specific case, as far as the "A" side is concerned, we connect TransportPCE to two xponder
+terminal devices at the netconf level :
+- XPDR-A1 is a 100GE transponder and is represented by XPDR-A1-XPDR1 node in *otn-topology*
+- SPDR-SA1 is an otn xponder that actually contains in its device configuration datastore two otn
+xponder nodes (the otn muxponder 10GE=>100G SPDR-SA1-XPDR1 and the otn switch 4x100GE => 4x100G SPDR-SA1-XPDR2)
+As represented on the bottom part of the figure, only one network port of XPDR-A1-XPDR1 is connected
+to the ROADM infrastructure, and only one network port of the otn muxponder is also attached to the
+ROADM infrastructure.
+Such network configuration will result in the TAPI *T0 - Multi-layer topology* abstraction as
+represented in the center of the figure. Let's notice that the otn switch (SPDR-SA1-XPDR2), not
+being attached to the ROADM infrastructure, is not abstracted.
+Moreover, 100GE transponder being connected, the TAPI *Transponder 100GE* topology will result in a
+single layer DSR node with only the two Owned Node Edge Ports representing the two 100GE client ports
+of respectively XPDR-A1-XPDR1 and XPDR-C1-XPDR1...
+
+
+**REST API** : *POST /restconf/operations/tapi-topology:get-topology-details*
+
+This request builds the TAPI *T0 - Multi-layer topology* abstraction with regard to the current
+state of *openroadm-topology* and *otn-topology* topologies stored in OpenDaylight datastores.
+
+**Sample JSON Data**
+
+.. code:: json
+
+ {
+ "tapi-topology:input": {
+ "tapi-topology:topology-id-or-name": "T0 - Multi-layer topology"
+ }
+ }
+
+This request builds the TAPI *Transponder 100GE* abstraction with regard to the current state of
+*openroadm-topology* and *otn-topology* topologies stored in OpenDaylight datastores.
+Its main interest is to simply and directly retrieve 100GE client ports of 100G Transponders that may
+be connected together, through a point-to-point 100GE service running over a wavelength.
+
+.. code:: json
+
+ {
+ "tapi-topology:input": {
+ "tapi-topology:topology-id-or-name": "Transponder 100GE"
+ }
+ }
+
+
+.. note::
+
+ As for the *T0 multi-layer* topology, only 100GE client port whose their associated 100G line
+ port is connected to Add/Drop nodes of the ROADM infrastructure are retrieved in order to
+ abstract only relevant information.
+
+This request builds the TAPI *T0 - Full Multi-layer* topology with respect to the information existing in
+the T-API topology context stored in OpenDaylight datastores.
+
+.. code:: json
+
+ {
+ "tapi-topology:input": {
+ "tapi-topology:topology-id-or-name": "T0 - Full Multi-layer topology"
+ }
+ }
+
+**REST API** : *POST /restconf/operations/tapi-topology:get-node-details*
+
+This request returns the information, stored in the Topology Context, of the corresponding T-API node.
+The user can provide, either the Uuid associated to the attribute or its name.
+
+**Sample JSON Data**
+
+.. code:: json
+
+ {
+ "tapi-topology:input": {
+ "tapi-topology:topology-id-or-name": "T0 - Full Multi-layer topology",
+ "tapi-topology:node-id-or-name": "ROADM-A1+PHOTONINC_MEDIA"
+ }
+ }
+
+**REST API** : *POST /restconf/operations/tapi-topology:get-node-edge-point-details*
+
+This request returns the information, stored in the Topology Context, of the corresponding T-API NEP.
+The user can provide, either the Uuid associated to the attribute or its name.
+
+**Sample JSON Data**
+
+.. code:: json
+
+ {
+ "tapi-topology:input": {
+ "tapi-topology:topology-id-or-name": "T0 - Full Multi-layer topology",
+ "tapi-topology:node-id-or-name": "ROADM-A1+PHOTONINC_MEDIA",
+ "tapi-topology:ep-id-or-name": "ROADM-A1+PHOTONINC_MEDIA+DEG1-TTP-TXRX"
+ }
+ }
+
+**REST API** : *POST /restconf/operations/tapi-topology:get-link-details*
+
+This request returns the information, stored in the Topology Context, of the corresponding T-API link.
+The user can provide, either the Uuid associated to the attribute or its name.
+
+**Sample JSON Data**
+
+.. code:: json
+
+ {
+ "tapi-topology:input": {
+ "tapi-topology:topology-id-or-name": "T0 - Full Multi-layer topology",
+ "tapi-topology:link-id-or-name": "ROADM-C1-DEG1-DEG1-TTP-TXRXtoROADM-A1-DEG2-DEG2-TTP-TXRX"
+ }
+ }
+
+T-API Connectivity & Common Services
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Phosphorus SR0 extends the T-API interface support by implementing the T-API connectivity Service.
+This interface enables a higher level controller or an orchestrator to request the creation of
+connectivity services as defined in the *tapi-connectivity* model. As it is necessary to indicate the
+two (or more) SIPs (or endpoints) of the connectivity service, the *tapi-common* model is implemented
+to retrieve from the datastore all the innformation related to the SIPs in the tapi-context.
+Current implementation of the connectivity service maps the *connectivity-request* into the appropriate
+*openroadm-service-create* and relies on the Service Handler to perform path calculation and configuration
+of devices. Results received from the PCE and the Rendererare mapped back into T-API to create the
+corresponding Connection End Points (CEPs) and Connections in the T-API Connectivity Context and store it
+in the datastore.
+
+This first implementation includes the creation of:
+
+- ROADM-to-ROADM tapi-connectivity service (MC connectivity service)
+- OTN tapi-connectivity services (OCh/OTU, OTSi/OTU & ODU connectivity services)
+- Ethernet tapi-connectivity services (DSR connectivity service)
+
+- RPC calls implemented
+
+ - create-connectivity-service
+
+ - get-connectivity-service-details
+
+ - get-connection-details
+
+ - delete-connectivity-service
+
+ - get-connection-end-point-details
+
+ - get-connectivity-service-list
+
+ - get-service-interface-point-details
+
+ - get-service-interface-point-list
+
+Creating a T-API Connectivity service
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Use the *tapi* interface to create any end-to-end connectivity service on a T-API based
+network. Two kind of end-to-end "optical" connectivity services are managed by TransportPCE T-API module:
+- 10GE service from client port to client port of two OTN Xponders (MUXPDR or SWITCH)
+- Media Channel (MC) connectivity service from client add/drop port (PP port of SRG) to
+client add/drop port of two ROADMs.
+
+As mentioned earlier, T-API module interfaces with the Service Handler to automatically invoke the
+*renderer* module to create all required tapi connections and cross-connection on each device
+supporting the service.
+
+Before creating a low-order OTN connectivity service (1GE or 10GE services terminating on
+client port of MUXPDR or SWITCH), the user must ensure that a high-order ODU4 container
+exists and has previously been configured (it means structured to support low-order otn services)
+to support low-order OTN containers.
+
+Thus, OTN connectivity service creation implies three steps:
+1. OTSi/OTU connectivity service from network port to network port of two OTN Xponders (MUXPDR or SWITCH in Photonic media layer)
+2. ODU connectivity service from network port to network port of two OTN Xponders (MUXPDR or SWITCH in DSR/ODU layer)
+3. 10GE connectivity service creation from client port to client port of two OTN Xponders (MUXPDR or SWITCH in DSR/ODU layer)
+
+The first step corresponds to the OCH-OTU4 service from network port to network port of OpenROADM.
+The corresponding T-API cross and top connections are created between the CEPs of the T-API nodes
+involved in each request.
+
+Additionally, an *MC connectivity service* could be created between two ROADMs to create an optical
+tunnel and reserve resources in advance. This kind of service corresponds to the OC service creation
+use case described earlier.
+
+The management of other OTN services through T-API (1GE-ODU0, 100GE...) is planned for future releases.
+
+Any-Connectivity service creation
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+As for the Service Creation described for OpenROADM, the initial steps are the same:
+
+- Connect netconf devices to the controller
+- Create XPDR-RDM links and configure RDM-to-RDM links (in openroadm topologies)
+
+Bidirectional T-API links between xpdr and rdm nodes must be created manually. To that end, use the
+following REST RPCs:
+
+From xpdr <--> rdm:
+^^^^^^^^^^^^^^^^^^^
+
+**REST API** : *POST /restconf/operations/transportpce-tapinetworkutils:init-xpdr-rdm-tapi-link*
+
+**Sample JSON Data**
+
+.. code:: json
+
+ {
+ "input": {
+ "xpdr-node": "<XPDR_OpenROADM_id>",
+ "network-tp": "<XPDR_TP_OpenROADM_id>",
+ "rdm-node": "<ROADM_OpenROADM_id>",
+ "add-drop-tp": "<ROADM_TP_OpenROADM_id>"
+ }
+ }
+
+Use the following REST RPC to invoke T-API module in order to create a bidirectional connectivity
+service between two devices. The network should be composed of two ROADMs and two Xponders (SWITCH or MUX)
+
+**REST API** : *POST /restconf/operations/tapi-connectivity:create-connectivity-service*
+
+**Sample JSON Data**
+
+.. code:: json
+
+ {
+ "tapi-connectivity:input": {
+ "tapi-connectivity:end-point": [
+ {
+ "tapi-connectivity:layer-protocol-name": "<Node_TAPI_Layer>",
+ "tapi-connectivity:service-interface-point": {
+ "tapi-connectivity:service-interface-point-uuid": "<SIP_UUID_of_NEP>"
+ },
+ "tapi-connectivity:administrative-state": "UNLOCKED",
+ "tapi-connectivity:operational-state": "ENABLED",
+ "tapi-connectivity:direction": "BIDIRECTIONAL",
+ "tapi-connectivity:role": "SYMMETRIC",
+ "tapi-connectivity:protection-role": "WORK",
+ "tapi-connectivity:local-id": "<OpenROADM node ID>",
+ "tapi-connectivity:name": [
+ {
+ "tapi-connectivity:value-name": "OpenROADM node id",
+ "tapi-connectivity:value": "<OpenROADM node ID>"
+ }
+ ]
+ },
+ {
+ "tapi-connectivity:layer-protocol-name": "<Node_TAPI_Layer>",
+ "tapi-connectivity:service-interface-point": {
+ "tapi-connectivity:service-interface-point-uuid": "<SIP_UUID_of_NEP>"
+ },
+ "tapi-connectivity:administrative-state": "UNLOCKED",
+ "tapi-connectivity:operational-state": "ENABLED",
+ "tapi-connectivity:direction": "BIDIRECTIONAL",
+ "tapi-connectivity:role": "SYMMETRIC",
+ "tapi-connectivity:protection-role": "WORK",
+ "tapi-connectivity:local-id": "<OpenROADM node ID>",
+ "tapi-connectivity:name": [
+ {
+ "tapi-connectivity:value-name": "OpenROADM node id",
+ "tapi-connectivity:value": "<OpenROADM node ID>"
+ }
+ ]
+ }
+ ],
+ "tapi-connectivity:connectivity-constraint": {
+ "tapi-connectivity:service-layer": "<TAPI_Service_Layer>",
+ "tapi-connectivity:service-type": "POINT_TO_POINT_CONNECTIVITY",
+ "tapi-connectivity:service-level": "Some service-level",
+ "tapi-connectivity:requested-capacity": {
+ "tapi-connectivity:total-size": {
+ "value": "<CAPACITY>",
+ "unit": "GB"
+ }
+ }
+ },
+ "tapi-connectivity:state": "Some state"
+ }
+ }
+
+As for the previous RPC, MC and OTSi correspond to PHOTONIC_MEDIA layer services,
+ODU to ODU layer services and 10GE/DSR to DSR layer services. This RPC invokes the
+*Service Handler* module to trigger the *PCE* to compute a path over the
+*otn-topology* that must contains ODU4 links with valid bandwidth parameters. Once the path is computed
+and validated, the T-API CEPs (associated with a NEP), cross connections and top connections will be created
+according to the service request and the topology objects inside the computed path. Then, the *renderer* and
+*OLM* are invoked to implement the end-to-end path into the devices and to update the status of the connections
+and connectivity service.
+
+.. note::
+ Refer to the "Unconstrained E2E Service Provisioning" use cases from T-API Reference Implementation to get
+ more details about the process of connectivity service creation
+
+Deleting a connectivity service
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Use the following REST RPC to invoke *TAPI* module in order to delete a given optical
+connectivity service.
+
+**REST API** : *POST /restconf/operations/tapi-connectivity:delete-connectivity-service*
+
+**Sample JSON Data**
+
+.. code:: json
+
+ {
+ "tapi-connectivity:input": {
+ "tapi-connectivity:service-id-or-name": "<Service_UUID_or_Name>"
+ }
+ }
+
+.. note::
+ Deleting OTN connectivity services implies proceeding in the reverse way to their creation. Thus, OTN
+ connectivity service deletion must respect the three following steps:
+ 1. delete first all 10GE services supported over any ODU4 to be deleted
+ 2. delete ODU4
+ 3. delete MC-OTSi supporting the just deleted ODU4
+
+T-API Notification Service
+~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+In future releases, the T-API notification service will be implemented. The objective will be to write and read
+T-API notifications stored in topics of a Kafka server as explained later in the odl-transportpce-nbinotifications
+section, but T-API based.
+
+
+odl-transportpce-dmaap-client
+-----------------------------
+
+This feature allows TransportPCE application to send notifications on ONAP Dmaap Message router
+following service request results.
+This feature listens on NBI notifications and sends the PublishNotificationService content to
+Dmaap on the topic "unauthenticated. TPCE" through a POST request on /events/unauthenticated.TPCE
+It uses Jackson to serialize the notification to JSON and jersey client to send the POST request.
+
+odl-transportpce-nbinotifications
+---------------------------------
+
+This feature allows TransportPCE application to write and read notifications stored in topics of a Kafka server.
+It is basically composed of two kinds of elements. First are the 'publishers' that are in charge of sending a notification to
+a Kafka server. To protect and only allow specific classes to send notifications, each publisher
+is dedicated to an authorized class.
+There are the 'subscribers' that are in charge of reading notifications from a Kafka server.
+So when the feature is called to write notification to a Kafka server, it will serialize the notification
+into JSON format and then will publish it in a topic of the server via a publisher.
+And when the feature is called to read notifications from a Kafka server, it will retrieve it from
+the topic of the server via a subscriber and will deserialize it.
+
+For now, when the REST RPC service-create is called to create a bidirectional end-to-end service,
+depending on the success or the fail of the creation, the feature will notify the result of
+the creation to a Kafka server. The topics that store these notifications are named after the connection type
+(service, infrastructure, roadm-line). For instance, if the RPC service-create is called to create an
+infrastructure connection, the service notifications related to this connection will be stored in
+the topic 'infrastructure'.
+
+The figure below shows an example of the application nbinotifications in order to notify the
+progress of a service creation.
+
+.. figure:: ./images/TransportPCE-nbinotifications-service-example.jpg
+ :alt: Example of service notifications using the feature nbinotifications in TransportPCE
+
+
+Depending on the status of the service creation, two kinds of notifications can be published
+to the topic 'service' of the Kafka server.
+
+If the service was correctly implemented, the following notification will be published :
+
+
+- **Service implemented !** : Indicates that the service was successfully implemented.
+ It also contains all information concerning the new service.
+
+
+Otherwise, this notification will be published :
+
+
+- **ServiceCreate failed ...** : Indicates that the process of service-create failed, and also contains
+ the failure cause.
+
+
+To retrieve these service notifications stored in the Kafka server :
+
+**REST API** : *POST /restconf/operations/nbi-notifications:get-notifications-process-service*
+
+**Sample JSON Data**
+
+.. code:: json
+
+ {
+ "input": {
+ "connection-type": "service",
+ "id-consumer": "consumer",
+ "group-id": "test"
+ }
+ }
+
+.. note::
+ The field 'connection-type' corresponds to the topic that stores the notifications.
+
+Another implementation of the notifications allows to notify any modification of operational state made about a service.
+So when a service breaks down or is restored, a notification alarming the new status will be sent to a Kafka Server.
+The topics that store these notifications in the Kafka server are also named after the connection type
+(service, infrastructure, roadm-line) accompanied of the string 'alarm'.
+
+To retrieve these alarm notifications stored in the Kafka server :
+
+**REST API** : *POST /restconf/operations/nbi-notifications:get-notifications-alarm-service*
+
+**Sample JSON Data**
+
+.. code:: json
+
+ {
+ "input": {
+ "connection-type": "infrastructure",
+ "id-consumer": "consumer",
+ "group-id": "test"
+ }
+ }
+
+.. note::
+ This sample is used to retrieve all the alarm notifications related to infrastructure services.
+
Help
----
Code Review / transportpce.git / blobdiff